Dana Golden, Aruna Balasubramanian, Niranjan Balasubramanian
arXiv 30 May 2026 · Econometrics
arXiv:2606.00811 · PDF · DOI · OpenAlex · Extracted main text
Data centers now account for 4.4% of United States electricity demand, yet the grid-level effectiveness of the renewable energy certificates (RECs) and power purchase agreements (PPAs) hyperscalers use to claim carbon neutrality remains unclear. We develop a game-theoretic model in which a data center operator chooses among RECs, PPAs, and behind-the-meter colocation while generators make entry decisions under endogenous financing costs. The model identifies a timing wedge -- the mismatch between consumption and credited renewable generation -- as a central mechanism through which AI demand degrades reliability, raises prices, and increases emissions even when RECs cover 100% of annual consumption. Colocation with storage addresses this wedge directly and induces the greatest renewable entry by eliminating generator revenue risk. We test these predictions by exploiting the staggered release of large language models as a natural experiment, using difference-in-differences on a novel dataset linking AI activity to local grid outcomes. AI demand significantly increases fossil generation, wholesale prices (up to 25% in treated PJM zones), and outage frequency (0.5--1 additional outages per year) near data centers, with impacts scaling in model size. Data centers with on-site generation exhibit a sign reversal in power-quality effects, consistent with the model's prediction that behind-the-meter capacity absorbs demand spikes. Counterfactual analyses show that edge inference, spatial reallocation, and colocated storage each substantially mitigate grid impacts, while REC-only strategies do not. Together, our results demonstrate that the externalities of AI to the grid are tightly coupled to procurement design and the spatial organization of data center infrastructure.
appendix boundary found by appendix_command · 70% of the source is main text. Read the extracted text to check this.
The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.
| Reference | Intensity | Mentions | Sections | Main text | |
|---|---|---|---|---|---|
| 1 | Callaway, Brantly and Sant'Anna, Pedro HC (2021) Difference-in-differences with multiple time periods | 0.737 | 4 | 3 | 50% |
| 2 | Schwartz, Roy and Dodge, Jesse and Smith, Noah A and Etzioni, Oren (2020) Green ai | 0.737 | 3 | 3 | 67% |
| 3 | Strubell, Emma and Ganesh, Ananya and McCallum, Andrew Energy and Policy Considerations for Deep Learning in NLP | 0.737 | 3 | 3 | 67% |
| 4 | Leonardo Nicoletti and Naureen Malik and Andre Tartar (2024) AI Needs So Much Power, It's Making Yours Worse | 0.644 | 2 | 2 | 100% |
| 5 | Cengiz, Doruk and Dube, Arindrajit and Lindner, Attila and Zipperer,… (2019) The effect of minimum wages on low-wage jobs | 0.644 | 2 | 2 | 100% |
| 6 | Deshpande, Manasi and Li, Yue (2019) Who is screened out? Application costs and the targeting of disability programs | 0.644 | 2 | 2 | 100% |
| 7 | De Feuillley, Christophe and Fontaine, Gérard (2023) Pass-Through in Residential Retail Electricity Competition: Evidence from Pennsylvania | 0.511 | 3 | 2 | 33% |
| 8 | MacKay, Alexander and Mercadal, Ignacia (2024) Deregulation, Market Power, and Prices: Evidence from the Electricity Sector | 0.511 | 2 | 2 | 50% |
| 9 | United States Environmental Protection Agency (2022) Power Sector Emissions Data (CAMPD): Clean Air Markets Program Data | 0.405 | 1 | 1 | 100% |
| 10 | Laughner, Theo and Heckman, Scott and Manzur, Asif and Sloop, Chris (2024) Analysis of Total Harmonic Distortion on the U.S. Electric Grid | 0.405 | 1 | 1 | 100% |
Showing the top 10 of 71 scored citations.